HP P6000 Replication Solutions Manager User Guide (T3680-96089, October 2012)

When replicating a host volume with PVLinks multipathing, replication manager jobs define only
one path in the replica. If necessary, you can use PVLinks to add paths to the replica after the job
is finished.
See CreateHostVolume and CreateHostVolumeGroup job commands.
Troubleshooting–Linux
Job with volume group remount fails–Linux
Problem
A job creates a replica of a Linux volume group and mounts it on a host. In subsequent steps, the
job unmounts the replica, and then attempts to remount it on another host. During the remount step,
the job fails.
Explanation / resolution
You cannot remount the replica of a Linux volume group because the context of the volume group
is lost after the first mount.
IMPORTANT: The job template named Replicate host volume(s), mount to a host, then to a different
host, contains steps to mount and remount a replica. Do not use the template with Linux volume
groups.
To resolve this issue:
1. If necessary, cleanup resources that were involved in the failed job.
2. Edit the job and remove the remount steps.
3. Create another job to mount the replica on a different enabled host.
Troubleshooting–Tru64 UNIX
Job with AdvFS replication fails–Tru64 UNIX
Problem
Job that replicates Tru64 UNIX AdvFS host volumes fails on job commands such as:
CreateHostVolumeFromDiskDevices
CreateHostVolumeGroup.
Messages in the replication manager event log indicate that the host volume could not be found
in the volume group.
In the host directory /var/adm/messages, a message similar to the following can appear:
Jun 5 18:50:44 rita vmunix: Found bad xor in sbm_total_free_space!
Corrupted SBM metadata file!
Jun 5 18:50:44 rita vmunix: An AdvFS domain panic has occurred due to
either a metadata write error or an internal inconsistency. This domain
is being rendered Jun 5 18:50:44 rita vmunix: Please refer to guidelines
in AdvFS Guide to File System Administration regarding what steps to
take to recover this domain.
Explanation / resolution
When a job is run, heavy I/O in any of the mounted filesets in the domain can lead to domain
panic, causing the job to fail.
Resolve any issues on the host, modify the job to use freezefs and thawfs utilities, and rerun the
job.
40 HP P6000 Replication Solutions Manager